A Statistical and Visual Analysis of Municipally-Maintained Trees in the City of Burlington, VT

1 Introduction

2 Visualization & Analysis

2.1 Data Cleaning

# read in csv file
trees <- read.csv("./Burlington_Trees.csv")
library("magrittr")

# separate Geo.Point column to latitude and longitude
# and convert to numeric variables
trees <- trees %>% 
         separate(Geo.Point, c("lat", "long"), ",") %>% 
         mutate(lat = as.numeric(lat),
                long = as.numeric(long))

# separate species column into genus, species
trees <- trees %>% 
         separate(species, c("genus", "species"), ",")

Warning: Expected 2 pieces. Missing pieces filled with `NA` in 4170 rows [1, 4, 5, 7, 9, 13, 20, 21, 22, 23, 24, 25, 27, 28, 29, 30, 31, 32, 33, 34, ...].

# remove the "spp" included as a place holder for genus
trees$genus <- gsub(" spp", "", trees$genus)

# correct the misspelling of mapl to maple
trees$genus <- gsub("mapl", "maple", trees$genus, fixed = TRUE)
trees$genus <- gsub("maplee", "maple", trees$genus)

# convert zeros in numeric columns to NA so they will not be included in graphs
# in this case, zero values are due to lack of information, not lack of value,
# so all were converted to NA values to be filtered out later
# repeat for blank values
trees[trees == 0] <- NA
trees[trees == ""] <- NA

# convert dates to better format
trees <- trees %>% 
  mutate(modified = as.yearmon(modified, "%m/%Y"))

# cleaned data frame
head(trees)

NA

2.2 Number of Trees by Species

```r
# number of trees by species
# histogram

<!-- rnb-source-end -->

<!-- rnb-chunk-end -->


<!-- rnb-text-begin -->



## Relationship between species abundance and Land Use


<!-- rnb-text-end -->


<!-- rnb-chunk-begin -->


<!-- rnb-source-begin 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 -->

```r
# landuse v species
# multiple bar chart
# see if certain types of trees are more common by business, residential, etc.

# format the data into a new data fram
landuse_by_genus <- xtabs(formula = ~ landuse + genus,
                       data = trees) %>% 
  prop.table(margin = "landuse") %>%  # calculates proportio
  
  # pipe into data.frame
  # Using data.frame() turns it from a table into a data frame
  data.frame() %>% 
  filter(Freq > .1) 

landuse_by_genus

landuse_species_bar <- ggplot(data = landuse_by_genus,
       mapping = aes(x = landuse,
                     fill = genus,
                     y = Freq)) + 
  
  # Now we need to tell geom_bar() to override it'd default choice of y
  # We do this with stat = "identity" (identity means "Use the y I gave you!")
  geom_bar(color = "black",
           stat = "identity",
           position = "fill") + 
  
  labs(y = "Proportion", 
       title = "Top Genus per Land Use Type",
       x = "Land Use Type")


landuse_species_bar

2.3 The Trees of Burlington

2.4 Appraisal Estimates

```r
# diameter v appraisal
# scatterplot probably

```

3 Machine Learning